[Jeff Morriss <jeff.morriss@xxxxxxxxxxx>]

Shaun Jackman wrote:
> This task seems to be a fundamentally broken idea. Why use a funky
> g_ntohl -- or any g_xxx function for that matter -- when there's a
> perfectly good, standard library function, namely ntohl?
>
> If a particular architecture doesn't provide ntohl, one may use
> autoconf to detect this case and provide ntohl. This case is the
> primary reason for using autoconf.

But why do all that autoconf work at all when glib has already done it 
and is guaranteed to provide a g_ntohl()?